Gonçalo Magno

User Management and Onboarding team, which owns how client organisations and their members are onboarded onto a wellbeing platform.

• Designed and built a public Organisations and Users API from inception to general availability (async FastAPI and Pydantic on ECS Fargate and Lambda, event-sourced over MongoDB and DynamoDB, Terraform); it is the platform's programmatic onboarding channel for enterprise clients across four regions.
• Delivered standards-first SCIM 2.0 provisioning (RFC 7643/7644) with PATCH, filtering, bulk operations and cross-region identity uniqueness; implemented and tested separate integrations with Microsoft Entra ID and Okta, and passed Microsoft's SCIM validator.
• Conceived and built, end to end, an AI assistant for customer-support agents, now in production: it streams answers from a Claude-based copilot with retrieval and MCP tools through a hardened FastAPI streaming gateway and an accessible React widget, with SSO, fail-closed access control, per-user rate limits, a spend kill switch and test-enforced PII/PHI-free logging.
• Hands-on in the Datadog-to-Dynatrace migration: OpenTelemetry over OTLP for containerised (ECS) workloads, W3C trace context propagated across Step Functions and the OneAgent layer for Lambda, with a dual run for a zero-gap cutover and the tagging standard encoded once and reused.
• Identified the cause of nearly half of the production CloudWatch bill (a monitoring integration polling metrics in every AWS region instead of the few in use) and got it fixed the next day.
• Raised the engineering baseline across the team's Python services: modern tooling (uv), pre-commit with Ruff and strict type checking, code and dependency scanning, keyless OIDC deployments and automated semantic-versioned releases on GitHub Actions and CircleCI.
• Built the team's shared AI tooling (an AI pull-request reviewer packaged as a reusable action, plus per-repository agent instructions and guardrails), making agentic development part of the everyday workflow.
• Helped design the next-generation onboarding pipeline and built two of its core components: strict Pydantic-based validation that rejects malformed client files before they reach the platform, and a diff engine that works out exactly which members to add, update or remove.
• Recognised on several occasions by leadership and peers; a regular presenter at the fortnightly engineering-wide demo, and an interviewer in technical hiring, helping bring several engineers onto the team.

Own consultancy (web-ops.tech), delivering backend engineering for scale-ups in fintech and web3, e-commerce and mobility.

• Zharta (on-chain lending, web3): built serverless services on AWS Lambda, DynamoDB and Terraform; introduced CloudWatch dashboards and alarms as code routed to Slack; removed deployment downtime by rolling out Lambda aliases; introduced dependency injection to make the services testable.
• Kencko (e-commerce): worked on Python services (FastAPI and Flask, SQLAlchemy) backed by PostgreSQL and SQLite, on Kubernetes; extracted logic from the monolith into a new FastAPI service; owned integrations with the Shopify GraphQL API, shipping carriers, Contentful and SendGrid.
• Rydes, now NAVIT (mobility): built payouts and a custom invoicing mechanism in NestJS and PostgreSQL on GKE; integrated digital banking and card payments, and CO₂ offsetting.

• Designed and built a device booking and provisioning system spanning many gaming platforms, operating systems and architectures, giving developers cloud-like access to physical test devices.

• Led a small team building an in-house provisioning automation framework in Python (Django).

Integrated master's degree, Portugal (BSc 2006, MSc 2008).